About This Study

As AI image generation becomes increasingly realistic, understanding how well humans can distinguish AI-generated images from real photographs is an important research question with broad implications for media literacy, content moderation, and public trust in digital media.

This platform presents participants with side-by-side image pairs — one AI-generated, one real — across a variety of content categories. By collecting response data at scale, we aim to measure human detection accuracy across categories, identify which types of content are easiest or hardest to distinguish, and track how perception shifts as generation technology evolves.

Each image pair is carefully curated to present images of similar subject matter, ensuring that participants are challenged to evaluate image authenticity rather than simply topic similarity. Response time is also recorded to support analysis of confidence and deliberation patterns.

Participation is fully anonymous. No account is required, and no personally identifiable information is collected. Sessions are tracked using a randomly generated identifier stored locally in your browser — this identifier cannot be linked back to any individual.

All collected response data — including which side was selected, response time, and whether the answer was correct — is published openly for the research community. By participating, you contribute to a public dataset that researchers can use to study human perception of AI-generated imagery.

You may clear your session at any time by clearing your browser's local storage. Doing so will generate a new anonymous session identifier on your next visit.